JP6953345B2 - Cargo damage information management system and cargo damage information management method used for it - Google Patents

Description

The present invention relates to a technique for managing cargo damage information.

In the system that manages cargo damage information, cargo damage information is managed at each scene of transportation for cargo, such as when carrying in the warehouse, changing the storage location in the warehouse, and when carrying out to the other party. Need to be done.

Patent Document 1 is a conventional technique related to this technical field. In Patent Document 1, a computer used by an administrator and a handheld terminal carried by an operator in a vehicle logistics management system capable of streamlining operations such as registration and search of a vehicle storage location in a terminal regarding vehicle logistics operations. The handheld terminal has a function of reading the first data including the vehicle ID from the vehicle medium and a function of acquiring positioning information by short-range communication with the GPS receiver. .. In addition, when the vehicle has a damaged part such as a scratch, the part is photographed by the photographing function of the handheld terminal, and the photographed data (damage information) is acquired and saved. Then, the point that the shooting data (damage information) is transferred to the center system together with the vehicle ID and the positioning information and stored in association with the database is disclosed.

Japanese Unexamined Patent Publication No. 2014-40317

In the case of cargo damage management, conventionally, as described in Patent Document 1, as soon as damage is found, the damaged part is photographed using a camera or the like, and the photographed data is stored and managed in a database. In this way, conventionally, as cargo damage information, the presence or absence of damage is confirmed by human judgment, a photograph is taken, and the damaged part, type, and level are manually determined for the photographed damage image. Was there.

In addition, the damage level is easily affected by the part of human judgment, the result may differ depending on the person who judges, and there is a problem that agreement may not be reached between the consignor and the consignee. ..

In view of the above problems, an object of the present invention is to reduce the work cost of the damage management work and to make the determination criteria fair.

In view of the above background technology and problems, the present invention is, for example, a cargo damage information management system for managing cargo damage information, which is a mobile terminal for photographing cargo and a mobile terminal for photographing cargo. It has a cargo damage information management device that manages cargo damage information together with the captured image and key information that identifies the cargo, and the cargo damage information management device automatically determines the damage site, type, and level as cargo damage information. It is configured to have an AI image analysis function.

According to the present invention, it is possible to reduce the work cost of the damage management work and realize the fairness of the determination criteria.

It is an overall block diagram of the cargo damage information management system in Example 1. FIG. It is a block diagram of the cargo damage information management apparatus and AI image analysis function construction apparatus in Example 1. FIG. It is a processing flow diagram which incorporated the AI image analysis of the cargo damage information management apparatus in Example 1. It is a processing flow diagram explaining the processing content of the AI image analysis processing in Example 1. FIG. It is a data structure diagram used in the cargo damage information management apparatus in Example 1. FIG. It is a determination flow chart of the part, the type, and the level of the damage by the AI image analysis processing in Example 1. FIG. It is a detection processing flow chart of the same image in Example 2. FIG. It is a processing flow diagram of the cargo damage information management apparatus which incorporated the same image detection processing in Example 2. FIG. It is a supplementary explanatory view of FIG.

Hereinafter, examples of the present invention will be described with reference to the drawings.

FIG. 1 is an overall configuration diagram of the cargo damage information management system in this embodiment. In FIG. 1, in the cargo damage information management system, a cargo damage information management device 1, a mobile terminal 2, a client terminal 3, and a customer system information management device 4 are connected by a network 5.
The cargo damage information management device 1 manages the damage information of the cargo, captures the damage image of the cargo with the mobile terminal 2 having a camera function at the timing of each scene of transportation for the cargo, and captures the damage image and the cargo. It is managed as cargo damage information together with the key information that identifies. As cargo damage information, it is possible to link the attributes of the damaged part, type (tear, dent, wetness, etc.) and level (large, medium, small, etc.), and to make a handwritten memo to indicate the damaged part. The client terminal 3 controls the output of a damage report based on the data registered in the cargo damage management. The customer system information management device 4 manages a database or the like of the customer's system, and provides information for performing analysis or the like for damage to cargo, which will be described later.

In this embodiment, the cargo damage information management device 1 will be described in detail later, but by incorporating AI (Artificial Intelligence) image analysis, the damage site, type, and level are automatically determined, and the damage management work is performed. Realize reduction of work cost and fairness of judgment criteria. Regarding the AI image analysis function, the AI image analysis function construction device 6 for constructing the AI image analysis engine is used separately from the cargo damage information management device 1.

FIG. 2 is a configuration diagram of the cargo damage information management device 1 and the AI image analysis function construction device 6 in this embodiment. In FIG. 2, the cargo damage information management device 1 has a CPU 10, a memory 11, and a storage device 12, and they are connected by a data bus 13. The storage device 12 stores a cargo damage information management program 121 and an AI image analysis program 122, and the cargo damage information management program 121 stored in the storage device 12 is expanded in the memory 11 and the CPU 10 is deployed in the cargo. By executing the damage information management program, cargo damage information management is executed by software processing. Further, by executing the AI image analysis program 122, the AI image analysis is executed by software processing.

Further, the AI image analysis function construction device 6 has a CPU 60, a memory 61, and a storage device 62, which are connected by a data bus 63. The AI image analysis engine 621 and the AI image analysis program 622 are stored in the storage device 62, and the AI image analysis engine reads the teacher data image file 7 into the AI image analysis engine 621 and learns the image so that the AI image analysis engine can perform an image. The analysis is performed and the AI image analysis program 622 is constructed. By reflecting the constructed AI image analysis program 622 in the cargo damage information management device 1, the AI image analysis function can be used in the cargo damage information management device 1. The reflection of the AI image analysis program 622 from the AI image analysis function construction device 6 to the cargo damage information management device 1 is, for example, installation via a storage medium (optical disk such as a CD, USB, etc.) or a network. The AI image analysis function can be enhanced by learning more teacher data. By re-incorporating the enhanced AI image analysis function into the already-operated cargo damage information management device 1, the AI image analysis function can be enhanced.

FIG. 3 is a processing flow diagram incorporating AI image analysis of the cargo damage information management device in this embodiment. In FIG. 3, by taking a picture of the cargo damage with a digital camera or the like in step S101, the photographed image is output in S102. The image taken in S102 is registered in S103. AI image analysis is executed on the registered image in S104 to detect the damage information of the cargo. The cargo damage information management device creates a post-damage image in S1052 from the original pre-damage image registered in S1051. Further, damage information is output in S1053. Damage information of the image before damage detection and the image after damage detection is registered in S106, and the database is stored in S107.

In the cargo damage information management device, it is possible to output a damage report in S1081 from the registered and stored information. As a result, the work cost of the damage management work can be reduced. Further, by linking with the customer system database in S109, it becomes possible to perform analysis in S1082, to see trends such as the cause of damage, and to consider countermeasures.

FIG. 4 is a processing flow diagram for explaining the processing contents of the AI image analysis S104 in FIG. In FIG. 4, the AI image analysis S104 determines the damaged part in S202, outputs the damaged part information in S203, determines the damage type in S204, outputs the damage type information in S205, and outputs the damage type information in S206, based on the captured image. The damage level is determined and the damage level information is output in S207. From the damage part information output in S203, an image after damage detection in which the damaged part is marked on the captured image is created in S1052, and the damage information is obtained from the damage type information output in S205 and the damage level information output in S207. Created as S1053. Further, the captured image is created as it is as an image before damage detection in S1051.

FIG. 5 is a data configuration diagram used in the cargo damage information management device in this embodiment. In FIG. 5, (A) shows a cargo information table, and "key information 1", "key information 2", and "key information 3" are provided as information for the user to identify the cargo. In the cargo damage information management device, a number uniquely identified internally is assigned as a "cargo key" to the key information set by the user. In addition, "damage level (large, medium, small, etc.)" and "damage content (blurring, dirt, dent, etc.)" can be held in the cargo information unit. (B) shows an image information table and holds photographed image information for each cargo information. A plurality of records of 0 or more can be held for one record of the cargo information table, and an "image key" for each photographed image is assigned to each "cargo key" specified in the cargo information table. (C) shows an image addition information table and holds damage information for each photographed image. It is possible to hold 0 or more records for one record in the image information table, and for each "cargo key" and "image key" specified in the image information table, "part" and "damage type (level)". To hold.

FIG. 6 is a flow diagram for determining the location, type, and level of damage caused by the AI image analysis process in this embodiment. In the image analysis process, the captured image to be targeted is analyzed, the damaged portion is determined in S401, and the presence or absence of damage is determined in S402. If it is determined that there is no damage, the process is terminated, and if it is determined that there is damage, the damage type / level is determined in S403. In the determination of the damage type / level of S403, the damage type (cardboard dent, pallet dent, etc.) and level (large, medium, small) are determined as shown in S4031 to S4034, instead of the process of determining the level after determining the type. Judgment is made in combination. It is determined by performing this up to S403n for the number n of combinations of damage type and level. After that, a record is created in the image addition information table that manages the damage level and the damage content for each image in S404 from the information of the damage type / level determination result in S403.

As described above, in this embodiment, the AI image analysis function automatically determines the damage site, type, and level, and links with the cargo damage information management device. As a result, the work cost of damage management work can be reduced, and by performing AI image analysis in determining the damage level, it is possible to make a fair judgment because the result is based on a certain judgment standard instead of a human opinion. Become.

In this embodiment, an example in which the same image detection process using the AI process is used in the cargo damage information management device to further reduce the work cost of the damage management work will be described.

FIG. 7 is a detection processing flow chart of the same image in this embodiment. In FIG. 7, the AI image is digitized in S502 based on the captured image in S102. Regarding the AI image quantification, the quantification is performed using the integrated value, the average value, etc. of the predetermined area based on the brightness value and the color information of each pixel. Since this process is a known process, detailed description thereof will be omitted. Then, in S503, a numerical value indicating the captured image is acquired. After that, the processing for a plurality of images to be compared is repeated between S504 and S509. For example, if the image to be compared is a captured image at the time of carrying out the cargo, the registered image at the time of loading is repeated in order to search for the same cargo as the captured image just taken from the registered image at the time of loading. As a repetitive process, the image to be compared is acquired in S505, and the image is digitized in the same process as in S502 in S506. Then, in S507, a numerical value indicating the image to be compared is acquired. In S508, the numerical values of the captured image and the comparison target image are compared, and if the threshold value is exceeded, it is determined that they are not the same, and the next comparison target image is processed. If it is within the threshold value, it is determined that the images are the same, and as the processing for the same image in S501, for example, processing such as flagging is performed, and the process proceeds to the processing of the next comparison target image.

FIG. 8 is a processing flow diagram of the cargo damage information management device incorporating the same image detection processing in this embodiment. In FIG. 8, the process is basically the same as that in FIG. 3, but the same image detection process is added to the process. In FIG. 8, the same processing as in FIG. 3 is designated by the same reference numerals, and the description thereof will be omitted. In FIG. 8, the same image detection process in the current process is performed in S603 from the captured image in S102. This process is the process of FIG. Then, S604 determines whether or not the same image is detected, and if there is, S605 performs duplicate registration error processing. This does not register duplicate images in S103 by displaying a message indicating that there are duplicate images. If the same image is not detected in S604, the captured image is registered in S103. As a result, duplicate registration of captured images can be realized.

Further, from the registered image of the captured image in S103, the same image detection process in the previous step is performed in S610. This process is the process of FIG. Then, in S611, it is determined whether or not the same image is detected in the previous process, and if there is, the damage information of the same image in the previous process registered in S612 is acquired. On the other hand, if the same image is not detected in the previous step in S611, the process ends.

In S613, the damage information of the image taken in S1053 is compared with the damage information of the same image in the previous process in S612, and if there is a change in damage by checking the type and level of damage, the damage information in S614. The update notification process is performed, and for example, a notification such as that the dirt is getting worse is given. If there is no change in damage, the process ends. In this way, in the processing of S610 to S614, for example, it can be determined whether or not there is a damage change between the time when the same cargo is carried in and the time when the same cargo is carried out.

FIG. 9 is a supplementary explanatory view of FIG. In FIG. 9, when an image is taken at the time of loading the cargo and the cargo B and D are damaged, and when an attempt is made to register the damaged cargo B and D as the photographed image at the time of carrying out, the cargo is processed by S605. B is processed as a duplicate registration error. Further, since the damage of the cargo D has changed due to the processing of S614 as compared with the time of carrying in, which is the previous process, the damage information update notification processing is performed.

As described above, according to this embodiment, by using the detection process of the same image in the cargo damage information management device, it is possible to suppress the occurrence of unnecessary work cost of the same image, and further reduce the work cost of the damage management work. can do.

Although the examples have been described above, the present invention is not limited to the above-mentioned examples, and various modifications are included. For example, the above-described embodiment has been described in detail in order to explain the present invention in an easy-to-understand manner, and is not necessarily limited to those having all the described configurations.

1: Freight damage information management device, 2: Mobile terminal, 3: Client terminal, 4: Customer system information management device, 5: Network, 6: AI image analysis function construction device, 10,60: CPU, 11,61: Memory , 12, 62: Storage device, 13, 63: Data 121: Cargo damage information management program, 122, 622: AI image analysis program, 621: AI image analysis engine

Claims (4)

A cargo damage information management system that manages cargo damage information.
A mobile terminal that shoots the cargo and
It has a cargo damage information management device that manages cargo damage information together with a photographed image taken from the mobile terminal and key information for identifying the cargo.
The cargo damage information management device inputs the photographed image to the AI image analysis engine learned by using the image file which is the teacher data, performs damage analysis processing, and performs the damage analysis process as the cargo damage information. levels have a automatically determining AI image analysis function,
The cargo damage information management device is
The cargo information table has a cargo key, which is key information for identifying the cargo, key information for the user to identify the cargo, and a damage level and damage content.
As the image information table, it is possible to hold a plurality of photographed image information of 0 or more for one record of the cargo information table, and the cargo key unit has an image key for each photographed image.
As the image addition information table, it is possible to hold 0 or more damage information for each of a plurality of captured images for one record in the image information table, and the cargo key and the image key unit designated in the image information table can be used. A cargo damage information management system characterized by retaining parts, damage types, and levels. A cargo damage information management system that manages cargo damage information.
A mobile terminal that shoots the cargo and
It has a cargo damage information management device that manages cargo damage information together with a photographed image taken from the mobile terminal and key information for identifying the cargo.
The cargo damage information management device inputs the photographed image to the AI image analysis engine learned by using the image file which is the teacher data, performs damage analysis processing, and performs the damage analysis process as the cargo damage information. It has an AI image analysis function that automatically determines the level, and has an AI image analysis function.
The cargo damage information management device has the same image detection function.
The same image detection function detects the same image as the captured image in the current process, and if there is the same image, the cargo damage information is not registered.
In addition, the same image detection function detects the captured image and the same image in the previous process, and if there is the same image, the damage changes by comparing the cargo damage information with the cargo damage information of the same image in the previous process. A cargo damage information management system characterized in that damage information update notification processing is performed if this is the case. This is a cargo damage information management method in which the CPU manages cargo damage information by software processing by executing a processing program expanded in memory.
When the CPU executes a cargo damage information management program as the processing program, it acquires a photographed image of the cargo and manages it as cargo damage information together with the photographed image and key information for identifying the cargo.
AI image analysis in which the CPU executes an AI image analysis program as the processing program to input the captured image to an AI image analysis engine learned using an image file that is teacher data and perform damage analysis processing. The function automatically determines the damage site, type, and level as the cargo damage information.
The cargo damage information is
The cargo information table has a cargo key, which is key information for identifying the cargo, key information for the user to identify the cargo, and a damage level and damage content.
As the image information table, it is possible to hold a plurality of photographed image information of 0 or more for one record of the cargo information table, and the cargo key unit has an image key for each photographed image.
As the image addition information table, it is possible to hold 0 or more damage information for each of a plurality of captured images for one record in the image information table, and the cargo key and the image key unit designated in the image information table can be used. , A cargo damage information management method characterized by maintaining the part, damage type, and level . This is a cargo damage information management method in which the CPU manages cargo damage information by software processing by executing a processing program expanded in memory.
When the CPU executes a cargo damage information management program as the processing program, it acquires a photographed image of the cargo and manages it as cargo damage information together with the photographed image and key information for identifying the cargo.
AI image analysis in which the CPU executes an AI image analysis program as the processing program to input the captured image to an AI image analysis engine learned using an image file that is teacher data and perform damage analysis processing. The function automatically determines the damage site, type, and level as the cargo damage information.
When the CPU executes the AI image analysis program, the captured image and the comparison target image are digitized into AI images.
When the CPU executes the cargo damage information management program, it has the same image detection function of comparing the numerical values of the captured image acquired by the AI image digitization and the comparison target image to determine whether they are the same image.
The same image detection function detects the same image as the captured image in the current process, and if there is the same image, the cargo damage information is not registered.
In addition, the same image detection function detects the captured image and the same image in the previous process, and if there is the same image, the damage changes by comparing the cargo damage information with the cargo damage information of the same image in the previous process. A cargo damage information management method characterized in that damage information update notification processing is performed when the damage information is updated .

Images